Subject: Re: sun-lamp CVS commits
To: None <pk@NetBSD.ORG>
From: Chris G Demetriou <Chris_G_Demetriou@BALVENIE.PDL.CS.CMU.EDU>
List: current-users
Date: 09/12/1995 20:16:39
> pk
> Mon Sep 11 15:41:26 EDT 1995
> Update of /a/cvsroot/src/sys/arch/sparc/dev
> In directory pain.lcs.mit.edu:/b/tmp/cvs-serv24080
> 
> Modified Files:
> 	esp.c 
> Log Message:
> We have no need for `esp_print()'; the only child we expect is a `scsibus',
> and there's nothing to say if it isn't there. Returning -1 is an error
> anyway as noted by Chuck Cranor.

Uh, what gets printed (nothing?) if it's not there?

It seems to me that printing _something_ is a _VERY_ good idea.  in
particular, i can imagine a case where a user has a 'complete' config
file (i.e. nothing wildcarded), then adds an esp for which the
SCSI bus isn't attached, and then is very confused by the fact that no
devices, etc., were found, but nothing was printed.



chris